Improve Management of Prison Unit Maintenance

The state should improve the management of prison maintenance and take advantage of opportunities for automation to reduce operating costs, particularly in controlling inventory.


Background
The Maintenance Division of the Texas Department of Criminal Justice (TDCJ) is responsible for ensuring that each prison s mechanical systems operate adequately. While most of the more than 500-member staff are assigned to individual prison units, each geographic region has staff to help as needed. Generally, the s taff assigned to each unit include one specialist in each of ten crafts, including: plumbing/pipe fitting, electrical, heating/ventilation and carpentry/painting. Additionally, some units may have boiler staff, laundry and kitchen repairmen, solid waste te chnicians, locksmiths , door mechanics, incinerator operators and landfill supervisors. 1~ Each unit is headed by a plant maintenance manager and an assistant unit maintenance supervisor. 2~

Given the physical proximity of many of the units, assigning specialized staff to each unit is unnecessarily costly and limits the staff s effectiveness. Many units have fewer than eight maintenance staff. Staff organized into such small groups with specialized functions are much more likely to remain idle.

Each work order issued for non-routine work is reviewed by each unit s maintenance staff to determine if the problem is correctible by one of their craftsmen. If they cannot correct the problem, the work order is sent to the area maintenance staff to examine the issue. If area maintenance finds that outside expertise is required, the work will be contracted. Routine work orders, such as preventive maintenance (PM) work, is also performed by each unit s maintenance staff.

To lower overall costs and increase inmates v ocational skills, inmate labor is used to help repair and maintain each unit. A review of a month of inmate and staff labor at three units (Hobby, Goree and Hughes) shows how well some units use inmate labor. In September 1992, 80 percent of the labor hour s for work orders, and 73 percent of PM work was performed by inmates at Hobby. Hughes and Goree were slightly less successful, with each using inmates for 71 percent of the work orders, and 62 percent and 60 percent of their PM work, respectively. 3~

While the inventory of supplies and parts is automated at the area or central level, it is not automated at the individual units. Once parts or supplies have been shipped to a unit, central inventory considers them consumed, when in fact the items may sit on the shelf at the unit for years. Accounting is conducted annually at each unit to set a dollar value to the inventory, but no attempt is made to determine which inventory is unnecessary or overstocked. Central maintenance stores have no idea what type of inventory is deposited at each unit. Consequently, one unit may order supplies when another unit has many of that exact item sitting in its warehouse.

The current manual system makes it almost impossible for staff to find out if a piece of equipment is under warranty when repairs are needed. Additionally, there is no way to be sure that all required PM is accomplished on a timely basis, or at all. The ov erall result is additional repair costs, some of which might have been covered under warranty. In addition, since warranties on equipment are sometimes only valid if routine maintenance is performed on schedule, a unit s inability to prove the maintenance was conducted at the appropriate intervals may invalidate the warranties.

The California Department of Corrections has found that depending solely on a centralized computer mainframe system is no longer efficient for a large correctional system. 4~ The move away from centralized mainframes also is occurring in Texas. The state s Department of Human Services changed to a local-site system composed of networks of microcomputers that communicate with a centralized mainframe. 5~

The focus of computerization in TDCJ, for all practical purposes, is almost exclusively on the mainframe computer housed in Huntsville, and currently, there are plans to increase the mainframe capacity. Using spreadsheets, word processing, database and spe cial purpose software to increase productivity is uncommon at TDCJ. Download capability through which authorized users could copy mainframe information into local microcomputer networks for statistical analysis, is not a current system feature. The system s needs are not being analyzed for savings that could be achieved by using vendor-supplied microcomputer systems and software. 6~

Various studies have been done on automated maintenance management systems. A.T. Kearney, Inc. estimated in a 1990 study the average benefits of computerized maintenance management systems at:
28 percent increase in maintenance, hourly and salaried productivity;
20 percent increase in equipment utilization;
19 percent decrease in maintenance and material costs; and
18 percent reduction in maintenance reordering inventory investment. 7~


Recommendations
A. All inventory and maintenance record keeping at Texas Department of Criminal Justice-Institutional Division (TDCJ-ID) should be automated.

This would significantly increase productivity in the work order control process and result in savings by reducing system-wide inventories. The system should be p urchased as soon as possible, and all new prison units should have the system on line when the unit opens.

B. The paint and carpentry staff should be centralized for prisons within a 15-mile radius. Other staff also should be shared, but a unit-by-unit evaluation of regular problem areas should be conducted by TDCJ management to ensure increased efficiency.

C. Where prison unit maintenance staff total less than eight, the assistant unit maintenance supervisor position should be eliminated.

Under the current organizational design, each unit maintenance supervisor has an assistant unit maintenance supervisor who does not have direct supervisory authority over any other staff. In units where the maintenance staff is eight or more, organization shoul d be redesigned so the assistant unit maintenance supervisors directly supervise all staff.

D. An independent study should determine computerization needs and strategies for TDCJ. This study should be performed before expanding mainframe computer hardware beyond what is needed to support immediate needs.

Microcomputers could be used to improve management in such areas as equipment use, analyzing maintenance and material costs, tracking institutional disturbances, cell allocation and population analysis, maintaining visitor lists, local key control, staffi ng in budgetary analysis, local area networks to support word processing, analyzing inmate data and tracking legislation to assess its affect on inmate population growth.


Implications
An automated maintenan ce system fulfills several functions. It maintains all PM records, warranty information and part numbers for each piece of equipment, and produces work orders detailing when the equipment needs maintenance and the steps which need to be accomplished to ful fill the PM requirements. A system also keeps records on non-routine work orders, by trade, building, equipment type, etc., to help track these activities. Generally, these systems have a time module built into them which, when activated, estimates the amo unt of time necessary to complete the task to assist in work scheduling.

An automated maintenance system would significantly reduce the chance that needed repair and maintenance work slip through the cracks and would establish a record of all maintenance for a particular piece of equipment. The system would maintain inventory records for parts control, monitor preventive maintenance, estimate time standards for maintenance jobs and organize a work order control process.

Expanding the inventory system t o individual prison units should require a minimal investment for additional software, since the hardware will have already been acquired as part of the automated maintenance system. Establishing an automated inventory system would require a detailed listi ng of all existing parts and supplies, and data entry of the information would be time consuming. However, using inmate labor to accomplish these tasks would render the costs negligible, while the benefits would be significant.

Centralizing some staff would increase productivity and result in less idle staff time. More effective use of time would make for a reduction in the the backlog of work orders, without an increase in costs.


Fiscal Impact
Acquiring necessary hardware and software to support these automated systems would cost an estimated $400,000 initially and about $60,000 per year to maintain. The internal development of software could substantially reduce costs. The initial start-up will be time consuming, but using inmate labor to develop the system will result in negligible costs. One maintenance automation system currently used by a community college district in Texas was estimated to cost less than $5,000 per site.

Based on available data, there is no way to calculate how much could be saved if the entire TDCJ inventory system were automated so that supplies could be easily identified and transferred among units. Based on a review of inventory, discussion with staff and a review of surplus and salvage sales, it is reasonable to assume that a 10 percent reduction in inventory level should be easily achieved with an automated inventory system. Given that the 1993 maintenance budget indicates that $10 million is to be spent on consumable supplies and that the division maintains an inventory val ued at $16 million, a 10 percent reduction in inventory could be obtained by a 16 percent reduction in the anticipated supply purchases in fiscal 1994-95. Supply items needed at one unit are frequently surplus stock at other units. The ability to identify and transfer items among units will permit this reduction in inventory to occur, thereby reducing purchases.

Eliminating the assistant unit maintenance supervisor position in units of less than eight will result in a savings to TDCJ of about $500,000, by reducing 21 positions. By centralizing just the paint and carpentry staff, the annual savings to TDCJ would be about $200,000, associated with the reduction of nine positions. Other maintenance functions should be reviewed for possible centralization.

To achieve the anticipated savings, the Legislature would need to reduce the agency appropriation by $3 million for the upcoming biennium, in association with the recommended reduction in staff and purchases for the Department s supplies inventory. Although 30 positions will be eliminated, those individuals should be able to relocate to one of the new prisons that are opening about every two or three months. The anticipated savings are net of associated costs necessary to acquire hardware and software to supp ort this automated system.

Savings to the
Fiscal General Revenue Change
Year Fund 001 in FTEs

1994 $1,320,000 -30
1995 1,660,000 -30
1996 860,000 -30
1997 860,000 -30
1998 860,000 -30



Endnotes
1 Texas Department of Criminal Justice-Institutional Division, Budget Document, LED360 (Huntsville, Texas, December 1992).
2 Texas Department of Criminal Justice-Institutional Division, Unit Maintenance Organization Charts (Huntsville, Texas, December 1992).
3 Texas Department of Criminal Justice-Institutional Division, Maintenance Report, September 1992 (Huntsville, Texas, September 1992).
4 Interview with Alan J. Duran, Chief of Information Systems Branch, California Department of Corrections (Sacramento, California,
December 10, 1992).
5 Texas Department of Human Services, Shared Resources, Advance Planning Document (Austin, Texas, August 1992); and Texas Department of Human Services, Information Systems Architecture (Austin, Texas, December 1992). (Draft).
6 Interview with Lonnie Eslick, Assistant Director for Data Processing and Communications, Institutional Division, Texas Department of Criminal Justice, Huntsville, Texas, November 18, 1992.
7 A. T. Kearney, Inc., Best of the Best: Maintenance System User Satisfaction and Impact on Maintenance Effectiveness (Chicago, Illinois, May 1990).